https://bugs.kde.org/show_bug.cgi?id=301470
--- Comment #5 from Martin Steigerwald <[email protected]> --- Bernd, thank you a lot for your comment and your attempt to help to move along with the threading code. From what I gathered so far at the KDE Randa Meetings the threading code is basically very optimized. Current developers do not see many ways to speed it up. The reasons for the slow display of message lists of large folder are elsewhere. One reason was that KMail supports a different font for each tag. So the font handling took a lot of CPU cycles. And there are some other reasons as well, likely calculating sizes which are now estimated and some other parts. Volker actually profiled this code already. Volker and Dan started to work on improvement KMail now that Akonadi from master shifted the bottleneck to KMail. And I already see some nice improvements. I can basically access a 43000 mail folder with threading in about 5 to 10 seconds and in another 10 seconds the threading basically completed. I sure bet further improvements are possible and developers work on that. There are still some bottlenecks in the display and as displaying is incrementally updated it has an influence on how fast the threading is completed. Also there are ideas to only do it for the newest say 1000-2000 mails in each folder and include more mails on demand or in background. For that there are plans to basically move the threading inside Akonadi and even cache the results. If you want to help to make it happen, I suggest you use kdesrc-build to build the latest code and ask in #kontact on Freenode or kdepim mailing list about it. I can help you with setting up your development environment in case you have questions or need hints to some documentation. For development questions you better contact Dan and Volker on mailing list or IRC. -- You are receiving this mail because: You are the assignee for the bug. _______________________________________________ Kdepim-bugs mailing list [email protected] https://mail.kde.org/mailman/listinfo/kdepim-bugs
